Re: ptrace() and kernel tasks
Zack Weinberg (zack@rabi.columbia.edu)
Tue, 11 May 1999 15:46:47 -0400
On Tue, 11 May 1999 19:44:03 +0100, David Woodhouse wrote:
>
>zack@rabi.columbia.edu said:
>> It might be sensible to do this for init as well; then the special
>> case for pid=1 can be removed from sys_ptrace(). OTOH, you might want
>> the core dump if init crashed.
>
>If init dies, the kernel dies immediately, and doesn't even give you the
>opportunity to sync the disks. So a core dump wouldn't actually flushed to
>disk anyway.
>
>I consider this behaviour to be a bug, but others think otherwise.
I would agree with you, except that the one time I had init die on me (it
got hit by the 2.2.5 out-of-memory bug) the system stayed up. And I believe
it did generate a core dump.
IIRC your problem had to do with sysrq killall - maybe that's different? in
particular, sysrq sync is handled by kflushd, if it's dead it won't do
anything.
(time to go experiment...)
zw
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/